Long Term Evolution (LTE) is defined as a new-generation wireless communications standard. An LTE network is a pure packet switched system, and does not support a conventional circuit switched service. Therefore, LTE can provide a voice service only by using an Internet Protocol (IP) voice. This undoubtedly brings about many new challenges. A user expects that voice quality in the LTE network can be as good as voice quality in a circuit switched network. However, in a voice communication scenario, a problem of one-way audio usually exists. The one-way audio means that one party of a call cannot hear a voice of the other party of the call. However, the one-way audio is a systematic problem, and relates to a core network element, a transmission network element, a wireless network element, a user terminal, and the like. For example, wireless signal coverage is relatively poor, a user terminal cannot synchronize with a serving cell due to a software/hardware problem of the user terminal, or a communication mechanism of a base station in a core network has a problem. Because a relatively large quantity of network elements are related, the one-way audio is more difficult to be resolved. Currently, finding of the one-way audio mainly depends on a customer complaint received after the problem of the one-way audio occurs. However, once the complaint is received, the one-way audio actually has existed for several hours or even more than one day, while a maintenance engineer cannot know in advance, and can only passively wait for the complaint. After receiving the complaint, the maintenance engineer can reproduce the one-way audio only by using a large quantity of manual dialing tests or the like. However, a reproducing probability is extremely low. This reduces efficiency of resolving the problem of the one-way audio, and affects quality of voice communication of the user.